1 / 24

Canberra, Australia

EFFECTS OF RANGING IN CIRCULAR POLARIZATION John Luck and Chris Moore John-luck@bigpond.com , cmoore@eos-aus.com. Canberra, Australia. MOTIVATION. It is all Dave Arnold’s fault! He predicts that Laser Ranging to uncoated targets in CIRCULAR POLARIZATION will produce:

Télécharger la présentation

Canberra, Australia

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. EFFECTS OF RANGING IN CIRCULAR POLARIZATION John Luck and Chris Moore John-luck@bigpond.com , cmoore@eos-aus.com Canberra, Australia

  2. MOTIVATION It is all Dave Arnold’s fault! He predicts that Laser Ranging to uncoated targets in CIRCULAR POLARIZATION will produce: • Shorter ranges, by maybe 4 mm. • Greater return rates. • Reduced scatter. He also predicts that, in LINEAR POLARIZATION, these quantities will be affected by the angle between the satellite velocity aberration vector and the direction of the polarization vector as it reaches the target. We have conducted ranging experiments at Stromlo SLR, to see whether these effects can be detected. Some preliminary results were presented at the Grasse Workshop 2007.

  3. MAIN EXPERIMENT • A Quarter-Wave-Plate (QWP) was used to convert from linear to circular polarization. • The QWP was mounted in a custom-built mechanical prototype “QWP INSERTER” and placed on the laser table in the Transmit Beam between the SHG and the Transmit/Receive Mirror. • By remote control, the QWP was alternately inserted into, and withdrawn from, the laser path during ranging. • This change of state occurred (generally) exactly on Normal Point bin boundaries. It took a couple of seconds, during which the laser was temporarily disabled. • Processing, after standard filtering and trend removal, examined the differences between Normal Points in the two states. • Have observed 2 LAGEOS-1, 2 LAGEOS-2 passes in Sep’07, and 1 ETS8 pass in Sep’08.

  4. QWP INSERTER : EARLY PROTOTYPE “OUT” of a laser path

  5. QWP INSERTER: “IN”on laser table

  6. QWP INSERTER: “OUT”

  7. PROCESSING INPUT DATA • FULL-RATE data from standard Stromlo processing – in which only the final filtered data are included. • NORMAL POINT data from standard Stromlo processing – but only as a check. • CPF PREDICTIONS from HTSI or JAXA. • QWP INSERTER Log, showing times of IN/OUT changes. NORMAL POINT CONSTRUCTION • Interpolate (Lagrange 8-point) CPF X,Y,Z (ITRF) iteratively to “Bounce” time w.r.t. Stromlo, and calculate predicted ranges. • Add refraction (Mendes-Pavlis) to predictions. • Fit trend polynomials of degrees 1– 9 through residuals, both IN and OUT together. Manually choose “best” fit degree, and re-fit, and add to preds. • “DEPARTURES” are residuals from complete predictions. • NPs formed from Average Departure in each NP bin, at Average Time.

  8. JUDICIOUS USE OF STATISTICS STATES: “IN” = CIRCULAR, “OUT” = LINEAR POLARIZATION NORMAL POINT ANALYSIS • Used only “ADJACENT PAIRS” of NPs, i.e. discarded NPs not having a neighbour of the opposite state on either side. This minimised external effects e.g. refraction anomalies, poor choice of trend function,….. • Calculated Mean of NPs (a few numbers) and RMS about that Mean, separately for each state. • Return Rates calculated only from “Adjacent Pairs” and best estimates of Number of Shots / NP bin. FULL RATE ANALYSIS • Calculated Mean and RMS of FR data (thousands), separately for each state.

  9. “DEPARTURES” FROM TREND FUNCTIONETS-8, 11 Sep’08

  10. JUDICIOUS USE OF STATISTICS (cont.) BOTH ANALYSES • Calculate Range Difference D between states: D = MEAN IN – MEAN OUT – 3.1 ps where 3.1 ps allows for thickness of QWP (1-way) got from catalogues: 1.8 mm. • Calculate Pooled RMS S using classical formula given in textbooks when discussing Student’s ‘t’ application, e.g. S=SQRT(s12/n2 + s22/n1) when n1, n2 large • Test H0: D = 0 against H1: D < 0 hence 1-tailed test, hence calc. Student’s “t” = D/S with n1 + n2 - 2 degrees of freedom. • Calculate Variance Ratio F using classical formula given in textbooks when discussing Fisher’s ‘F’ applications: F = s12(IN)/s22(OUT) with n1-1, n2-1 degrees of freedom • Test H0: F = 1 against H1: F < 1 hence 1-tailed test. (This test has very weak power when F ~ 1.)

  11. JUDICIOUS USE OF STATISTICS (cont.) • Comparison of Return Rates not statistically tested. But B = [Acc.Returns/Num shots](IN) / [Acc.Returns/Num shots](OUT) using only numbers in Adjacent Pairs of NP bins. • Confidence Level in following tables is [1 – Pr{accepting H1 when H0 is TRUE}] x 100 %. CONTRARY COMPARISON • An ordinary LAGEOS pass in which the QWP was never inserted, was analysed to get a feeling for whether the procedure is valid. A dummy QWP log was constructed with random assignment of IN or OUT among NPs. • In this pass, it is reasonable to suppose that the Null Hypothesis H0: equality between “IN” (Circular) and “OUT” (Linear) will be accepted in all tests.

  12. EXAMPLE: ETS-8 on 11 Sep’08

  13. RESULTS

  14. CONCLUSIONS – MAIN EXPERIMENT • There is statistical evidence that : RANGE MEASUREMENTS ARE SHORTER IN CIRCULAR POLARIZATION than in Linear seen more clearly in Full Rate analysis, although Normal Point analysis comes tantalizingly close to the same conclusion. • There is : NO EVIDENCE THAT RMS SCATTER IS SMALLER IN CIRCULAR POLARIZATION than in Linear. (Using Fisher’s F-test with Full Rate data is meaningless!) • There is some evidence to suggest that, perhaps: RETURN RATES ARE GREATER IN CIRCULAR POLARIZATION than in Linear.

  15. ELECTRIC VECTOR ORIENTATION • The orientation of the Electric Vector (EV) of Linearly Polarized Light has been mapped through the Stromlo SLR Coude path. • The p- (parallel to plane of incidence) and s- (perpendicular) amplitude reflectances of each (plane) reflecting surface are taken into account, using data from handbooks, catalogues and text-books. • Initial EV is vertically upwards, after the SHG (confirmed by EOS opticians), and is assigned a magnitude 1. • Mapping ignores curved surfaces such as lenses, beam expanders and telescope primary and secondary mirrors (none of which are “turning mirrors”) so the output beam is as transmitted from the Tertiary Mirror.

  16. ELECTRIC VECTOR ORIENTATION • Calculations are done in local E,N,U coordinates and transformed to “Sky” coordinates after the tertiary mirror: “Sky” = {Magnitude, & Position Angle w.r.t Vertical Circle}. (E is the transmitted Electric Vector in E,N,U components)

  17. Example: EV POSITION ANGLE (ψ) Enhanced Aluminium (p = .967, s = -.990) for Coude mirrors, perfect for laser table mirrors.

  18. Extreme Example: EV TRANSMITTED ENERGY • Enhanced Aluminium (as in previous slide). • This sort of variation was measured at SGF 7840 last year and reported at Grasse • When reflectances are equal or perfect, the variations disappear.

  19. VELOCITY ABERRATION VECTOR V Satellite velocity vector, numerically differentiated in (rotating) ITRF from CPF predicted position vectors. X Satellite geocentric position vector from CPF in ITRF s Station geocentric position vector in ITRF Ω Earth Rotation vector R(φ,λ) Rotation from ITRF to Topocentric E, N, U a Velocity aberration vector in E,N,U: a ={ R(φ,λ)[V + Ω x (X – s)]} as a unit vector E Electric Vector (unit) expressed in E,N,U components If θ is the angle between the plane of polarization and the velocity aberration at the satellite – the “ARNOLD ANGLE”, then: θ = cos-1(a.b)

  20. ARNOLD ANGLE • Angle between polarization and aberration vectors calculated for one of the QWP LAGEOS-2 passes. It is not quite smooth ? • For the ETS-8 pass, it is very nearly constant, at 122o.8 • (but I have not yet worked out what to do with this information . . .)

  21. PERILS OF NUMERICAL DIFFERENTIATION • Essential to get BEST POSSIBLE TREND CURVE in this type of analysis, when looking for very small signals. • Try to fit a physically meaningful function BEFORE fitting the final Trend Polynomial so that the polynomial is of lowest reasonable degree. • Tried to fit empirical RANGE BIAS and TIME BIAS: Obs(i) – Pred(i) = RB + Range_Rate(i)*TB • Tried to fit OSCULATING KEPLER ELEMENTS to the predictions: a, e, i, Ω, ω, t0 from X,Y,Z, Xdot, Ydot, Zdot • They revealed DISCONTINUITIES in the RATES using Numerical Differentiation of CPF Positions (Lagr. Order 8)

  22. RATE DISCONTINUITIES • To make discontinuities visible, removed a line from the rates and took first differences. • Rates calculated at time of each FR point in a STARLETTE pass.

  23. RATES OF CPF POSITIONS • CONCLUSION: RATE DISCONTINUITIES EXIST IN CPF PREDICTIONS, which spoil efforts to find the best possible Trend Function.

  24. SUMMARY • There is statistical evidence from ranging observations to support the prediction that RANGE MEASUREMENTS ARE SHORTER IN CIRCULAR POLARIZATION THAN IN LINEAR, by 1 mm or so. • There is tantalizing evidence that RETURN RATES are GREATER in CIRCULAR POLARIZATION. • There is little evidence that SCATTER is reduced. • Imperfect p- and s-reflectances in Coude Path affect transmitted energy systematically in Az & El in LINEAR POLARIZATION. • Angle between Velocity Aberration and Electric Vectorsare readily calculated. • CPF PREDICTIONS are subject to RATE DISCONTINUITIES.

More Related